随笔分类 - read
以摘录为主，记录一些有可能会用到的方法或灵感
摘要：## pre title: FontTransformer: Few-shot High-resolution Chinese Glyph Image Synthesis via Stacked Transformers from: [Pattern Recognition 2023](https:
阅读全文
摘要：pre title: FontRL: Chinese Font Synthesis via Deep Reinforcement Learning accepted: AAAI 2021 paper: https://ojs.aaai.org/index.php/AAAI/article/view/
阅读全文
摘要：pre title: Continuous control with deep reinforcement learning accepted: ICLR 2016 paper: https://arxiv.org/abs/1509.02971v6 code: none ref: https://z
阅读全文
摘要：pre title: Instance Segmentation for Chinese Character Stroke Extraction, Datasets and Benchmarks accepted: arxiv paper: http://export.arxiv.org/abs/2
阅读全文
摘要：pre title: LANIT: Language-Driven Image-to-Image Translation for Unlabeled Data accepted: arxiv paper: arxiv code: https://github.com/KU-CVLAB/LANIT 亮
阅读全文
摘要：pre title: Rethinking the Truly Unsupervised Image-to-Image Translation accepted: ICCV 2021 paper: arxiv | ICCV code: https://github.com/clovaai/tunit
阅读全文
摘要：pre title: Unpaired Image Translation via Vector Symbolic Architectures accepted: ECCV2022 paper: arxiv code: https://github.com/facebookresearch/vsai
阅读全文
摘要：1. pre title: Few-shot Font Generation with Weakly Supervised Localized Representations accepted: TPAMI 2022 (extension of Few-shot Font Generation wi
阅读全文
摘要：PNAS最新研究：81%解题率，神经网络 Codex 推开高等数学世界大门 https://mp.weixin.qq.com/s/Eu0Iuy-bXJCq8JtgR-3yXA 论文：https://www.pnas.org/doi/10.1073/pnas.2123433119 在这项新研究中，研究
阅读全文
摘要：1. pre title: FontNet: Closing the gap to font designer performance in font synthesis accepted: AI4CC 2022 (AI for Content Creation, CVPR Workshop) pa
阅读全文
摘要：1. pre title: Look Closer to Supervise Better: One-Shot Font Generation via Component-Based Discriminator paper: https://arxiv.org/abs/2205.00146 code
阅读全文
摘要：MLP-Mixer: An all-MLP Architecture for Vision pre ref: https://zhuanlan.zhihu.com/p/369970953 ref: https://zhuanlan.zhihu.com/p/373070779 ref: https:/
阅读全文
摘要：pre ref: https://zhuanlan.zhihu.com/p/542389717 paper: https://arxiv.org/abs/2205.09965 code: https://github.com/tlc121/FsFont 阅读论文 《Few-Shot Font Gen
阅读全文
摘要：pre title: DG-Font: Deformable Generative Networks for Unsupervised Font Generation accepted: CVPR 2021 paper: https://openaccess.thecvf.com/content/C
阅读全文
摘要：1. 原作 论文：Multiple Heads are Better than One: Few-shot Font Generation with Multiple Localized Experts 代码：https://github.com/clovaai/mxfont 2. 摘要 小样本字体
阅读全文
摘要：Pix2Seq：谷歌大脑提出 CV 任务统一接口！ 以自然语言为输出的任务，比如image captioning、visual question answering这类任务，天然可以转化为生成text token sequence。但模型的输出形式还存在很多其他的形式，例如bounding box、
阅读全文
摘要：GANs N' Roses: Stable, Controllable, Diverse Image to Image Translation (works for videos too!) 知乎分析：论文阅读：GANs N’ Roses: Stable, Controllable, Diverse
阅读全文
摘要：2022-6-28 11:47:49 Paddle 生成对抗网络七日打卡营 GAN发展介绍 DCGAN 顾名思义，DCGAN主要讨论 CNN 与 GAN 如何结合使用并给出了一系列建议。由于卷积神经网络(Convolutional neural network, CNN)比MLP有更强的拟合与表达能
阅读全文
摘要：从零带你入门stylegan~stylegan3的技术细节 https://blog.csdn.net/xjm850552586/article/details/122027019 Mapping network: 用于将latent code z 转换成为 w Synthesis network:
阅读全文
摘要：0. 原作 论文：XMP-Font: Self-Supervised Cross-Modality Pre-training for Few-Shot Font Generation 代码：https://github.com/lfy523/XMP-Font （截止发文时还未公开代码，仅有仓库） 1
阅读全文
